skills.rest
silverstein
silversteinCOMMUNITY
Explore all skills by this author →
Website𝕏@matsilversteinGitHub

geniml

Produces Tunring-style songwriting sprints with customizable durations and tempos.

21|2|Updated Dec 8, 2025
Search Tags:#embeddings#region2vec#bed files#metadata search#genomic intervals#single-cell atac-seq#consensus peaks
One-click install
npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill geniml-silverstein
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: geniml
Source: https://github.com/silverstein/claude-scientific-skills-desktop/tree/main/corpus/geniml
Command: npx skills add https://github.com/silverstein/claude-scientific-skills-desktop --skill geniml-silverstein

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Genomic interval analysis is slow and hard to search because BED files and peak regions lack compact, comparable representations for ML and similarity queries, especially across experiments and single cells.

Core Features & Use Cases

  • Region2Vec embeddings for BED similarity: Learn unsupervised vector representations of genomic regions suitable for clustering, dimensionality reduction, and downstream ML features.
  • BEDspace joint embeddings for regions and metadata: Train shared embeddings that support metadata-aware queries across region sets and labels using a StarSpace-style workflow.
  • scEmbed for single-cell ATAC-seq: Produce cell-level embeddings from scATAC-seq peak coordinates for clustering and visualization in scanpy-compatible formats.
  • Consensus peak “universe” building: Create standardized peak vocabularies from BED collections using CC/CCF/ML/HMM methods to improve tokenization consistency.
  • Utilities for tokenization, caching, evaluation, and null models: Support practical ML workflows via tokenization helpers, BED caching, randomization (BEDshift), and embedding quality evaluation.

Quick Start

Use geniml to build a consensus universe from your combined BED files and then train Region2Vec embeddings for region similarity search.

Frequently Asked Questions about geniml

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate embeddings from BED files for genomic interval similarity search?▼

You can generate embeddings from BED files by applying Region2Vec to learn unsupervised vector representations of genomic intervals, enabling similarity search, clustering, and dimensionality reduction across bulk region datasets.

Can I create cell-level embeddings from single-cell ATAC-seq peak data?▼

Yes, you can produce cell-level embeddings from single-cell ATAC-seq peak coordinates using scEmbed, which outputs scanpy-compatible formats for clustering and visualization.

What is the best way to build a standardized peak vocabulary for tokenizing genomic regions?▼

The best way to build a standardized peak vocabulary is constructing a consensus peak universe from BED collections using CC, CCF, ML, or HMM methods to ensure consistent tokenization across experiments.

How do I train joint embeddings that combine genomic regions and metadata labels?▼

You can train joint embeddings combining genomic regions and metadata labels using BEDspace, which applies a StarSpace-style workflow to support metadata-aware queries across region sets.

Do I need to construct a consensus peak universe before training region embeddings?▼

Yes, building a consensus peak universe is required as a standardized token vocabulary to drive the BED-to-token pipeline before applying Region2Vec or joint embedding training.

What utilities are available for evaluating genomic interval embedding quality and generating null models?▼

Available utilities include tokenization helpers, BED caching, BEDshift randomization for null models, and evaluation-oriented outputs for region- and label-aware queries to assess embedding quality.

Related Skills

Explore More Education & Research Skills→
LeonChaoXLeonChaoX
783

geniml

Build machine learning embeddings from genomic interval BED files.

Community
Advanced
jonnabiojonnabio
6

scanpy

Automate single-cell RNA-seq analysis with scanpy from QC to visualization.

Community
Advanced
Estrella-231Estrella-231

geniml

Generate machine learning embeddings for genomic intervals from BED files.

Community
Intermediate
crazymsncrazymsn
22

geniml

Generate genomic interval embeddings from BED files using Region2Vec.

Community
Intermediate